The Difference Between Narrow AI and General AI
Innovating Tasks ; Solving Minds
Artificial Intelligence (AI) has advanced rapidly in recent years, with systems achieving and exceeding human performance on specific, limited tasks like sensing images, translating languages, driving vehicles, playing strategic games, and more. For example, AI can now detect objects in photos with superhuman accuracy, translate between dozens of languages instantly, defeat world champions at chess and Go, and navigate roads autonomously in vehicles.
However, we have not yet achieved general artificial intelligence—an AI with the broad, versatile intelligence that humans possess. General AI remains limited to science fiction rather than technological reality.
There are two major types of AI: narrow AI and general AI.
Narrow AI
Narrow AI, also known as Weak AI, is AI that is focused on a specific, limited task. Narrow AI systems are trained to perform a single job and they do it very well within a limited scope or domain. Examples of Narrow AI include:
Virtual assistants like Siri, Alexa and Cortana that can understand speech and respond to basic questions or commands. They are programmed with natural language processing capabilities to have conversations, but only within a limited range.
Image recognition software that can detect and label objects, faces or scenes in pictures. These systems are trained on thousands of images to identify specific images but they cannot generalize their learning to new types of images or tasks.
Self-driving cars that can sense the environment around them and navigate roads. They are trained to do the specific task of driving a vehicle but do not have a broad, general intelligence.
AI that plays strategic games like chess and Go. These systems are programmed with complex algorithms to evaluate positions and strategies to beat opponents at specific games. They do not have a generalized problem-solving intelligence.
General AI
General AI, also known as Strong AI, refers to machine intelligence that matches or exceeds human intelligence. Systems with general AI would have the broad, flexible thinking needed for complex thought, adaptive problem solving, social/emotional cognition and multidimensional learning—with perhaps superior speed, memory, perception and computational abilities compared to humans.
However, general AI of this caliber remains science fiction and is not technologically achieved yet. We have not built machines that replicate the general, multifaceted intelligence evident in human thinking, behavior and understanding. General AI could greatly amplify human potential, but it also introduces existential risks that require proactive safeguarding of its development and application.
As AI continues progressing rapidly with both narrow and general promises on the horizon, a clear understanding of their distinctions is crucial to reaping rewards and avoiding regret. With proper oversight, narrow AI can positively transform many areas of life, while general AI may radically reshape human progress itself—if ever realized at that level. But both types of AI have the potential for good and ill, expansion and destruction.
Endnote: Narrow AI and general AI differ greatly in scope, capability, flexibility, and potential impact. Narrow AI has achieved a lot but remains narrow in scope, while general AI could potentially match or exceed human intelligence in a far more multifaceted and flexible manner if ever achieved.
Both represent exciting yet uncertain possibilities depending on how they progress.